Apple to deploy MCP support for powerful AI experiences

It is a universally acknowledged truth that Apple has fallen behind in artificial intelligence, even though millions of its products can already run most cloud-based models as well as its own on-device AI, presently in the form of Apple Intelligence.

Which is to say: never believe something until it becomes true. You see, despite the story, the facts are that Apple has successfully developed, built, and distributed an ecosystem that is now just waiting for the AI software to be locked into place. 

The hardware is ready, next comes the code.

Up, up, and AI

While delayed, the code is coming. 9to5Mac claims the first developer beta of iOS 26.1 includes code to “suggest” Apple is working toward implementing support for the open-source Model Context Protocol (MCP) developed by Anthropic in its operating systems. 

(MCP is widely used by other AI companies, including OpenAI and Google.)

If true, this potentially unlocks the kind of useful, human-centred AI tasks we believe Apple dreams of — contextually aware solutions that can answer complex questions by interacting with one, two, or even more apps. 

What does this mean?

So, what does this mean to you, your users, and your business? 

The facts is, right now we can’t know precisely,. But we can make an educated guess. Hustle over to GitHub and you’ll find a collection of independently-developed Apple-native tools for MCP. The developer says these tools give Macs “superpowers.” If true, these powers include:

Create, search through and find Notes with spoken interrogations. 

Find your contact details without endless scrolling.

Send emails, including attachments, using voice.

Search emails the same, too.

Create and search for events, listing upcoming event, and reminders.

It also lets you string commands together in unique ways, such as, “Read my conference notes, find contacts for the people I met, and send them a thank you message.” It’s not much of a stretch to imagine it being possible for on-device AI to take your company data to handle really complex customer-facing and back-end tasks. 

Making your data work for you

“Which seat does this passenger like most?”

“What meal do they enjoy?”

“How many milling machines are at this location, when are they due to be serviced, and when did the last repair take place?” 

These are all examples of the kinds of tasks that could be made easy. Of course, the beauty of AI extends far beyond being provided a service manual, booking collection and delivery, and ordering spare parts — all within one swift spoken Siri interaction. 

In brief, MCP enables your device to interact in unique ways with multiple apps to get sometimes very complex tasks done, including all those tasks you keep meaning to do, but never get round to — for example, organizing your contacts database or building a better filing system in Notes.

App Intents+

Apple has also shown us some of what it plans in the form of App Intents and Shortcuts integrations. These offer similar sounding sets of tools to the above, and Apple has explained that App Intents for developers: “Provides functionality to deeply integrate your app’s actions and content with system experiences across platforms, including Siri, Spotlight, widgets, controls and more.” Add App Intents to Shortcuts and add a sprinkling of MCP and you can see what Apple is probably striving for.

The myth, the legend

Apple also wants to ensure that all your data stays on your device. The idea here is that despite the AI functionality, the information will remain private, unless you approve its use for more complex tasks than the device can handle. Given human nature, it matters that this data does remain private, given that one recent study showed 38% of US employees have shared proprietary product information with AI tools online.

That’s why it matters that we already know a MacBook Pro can run an LLM model natively on the device and that the new processor inside the latest iPhone Pro delivers MacBook Pro levels of computational power. 

(Sensibly, it makes sense to build devices that can run their own models, as doing so distributes capex costs while also likely reducing water and energy costs, which will become even more important as those costs peak.)

Do you think that power is just there for bragging rights? Of course it isn’t. From where I sit, the universally acknowledged “truth” — that Apple is behind on AI — will soon become yet another fairy tale written to help stressed-out competitors sleep better, like gingerbread houses, horses with wings, and little green men.

You can follow me on social media! Join me on BlueSky,  LinkedIn, and Mastodon.